def compute_prior(self):
self.prior = GaussianLOTHypothesis.compute_prior(self)
self.prior += sum(
map(lambda x: normlogpdf(x, 0.0, CONSTANT_SD),
self.CONSTANT_VALUES))
self.posterior_score = self.prior + self.likelihood
return self.prior
def toMaximize(fit_params):
self.CONSTANT_VALUES = fit_params.tolist() # set these
# And return the original likelihood, which by get_function_responses above uses this
constant_prior = sum(
map(lambda x: normlogpdf(x, 0.0, CONSTANT_SD),
self.CONSTANT_VALUES))
return -(GaussianLOTHypothesis.compute_likelihood(self, data) +
constant_prior)
def to_maximize(fit_params):
self.parameters = fit_params.tolist() # set these
# And return the original likelihood, which by get_function_responses above uses this
constant_prior = sum(
map(lambda x: normlogpdf(x, 0.0, self.constant_sd),
self.parameters))
return -(LOTHypothesis.compute_likelihood(self, data) +
constant_prior)
def compute_single_likelihood(self, datum):
""" Compute the likelihood with a Gaussian. Wraps to avoid nan"""
ret = normlogpdf(self(*datum.input), datum.output, datum.ll_sd)
if isnan(ret):
return -Infinity
else:
return ret
def compute_single_likelihood(self, datum):
""" Compute the likelihood with a Gaussian. Wraps to avoid nan"""
ret = normlogpdf(self(*datum.input), datum.output, datum.ll_sd)
if isnan(ret):
return -Infinity
else:
return ret
def compute_single_likelihood(self, datum):
v = self(*datum.input)
ret = sum([normlogpdf(vi, di, datum.ll_sd) for vi, di in zip(v, datum.output)])
if isnan(ret):
return -Infinity
else:
return ret
def generate(self, x='*USE_START*', d=0):
"""
RealValueGrammar.generate may create gaussians or uniforms when given "*gaussian*" and "*uniform*" as the nonterminal type.
Otherwise, this is identical to LOTlib.Grammar
"""
if x == '*USE_START*':
x = self.start
if x=='*gaussian*':
# TODO: HIGHLY EXPERIMENTAL!!
# Wow this is really terrible for mixing...
v = np.random.normal()
gp = normlogpdf(v, 0.0, 1.0)
return FunctionNode(returntype=x, name=str(v), args=None, generation_probability=gp, ruleid=0, resample_p=CONSTANT_RESAMPLE_P ) ##TODO: FIX THE ruleid
elif x=='*uniform*':
v = np.random.rand()
gp = 0.0
return FunctionNode(returntype=x, name=str(v), args=None, generation_probability=gp, ruleid=0, resample_p=CONSTANT_RESAMPLE_P ) ##TODO: FIX THE ruleid
else:
# Else call normal generation
Grammar.generate(self,x,d=d)
def to_maximize(fit_params):
self.parameters = fit_params.tolist() # set these
# And return the original likelihood, which by get_function_responses above uses this
constant_prior = sum(map(lambda x: normlogpdf(x,0.0,self.constant_sd), self.parameters))
return -(LOTHypothesis.compute_likelihood(self, data) + constant_prior)
def compute_prior(self):
# Add together the LOT prior and the constant prior, here just a gaussian
return LOTHypothesis.compute_prior(self) +\
sum(map(lambda x: normlogpdf(x,0.0,self.constant_sd), self.parameters))
def to_maximize(fit_params):
self.CONSTANT_VALUES = fit_params.tolist() # set these
# And return the original likelihood, which by get_function_responses above uses this
constant_prior = sum(map(lambda x: normlogpdf(x,0.0,CONSTANT_SD), self.CONSTANT_VALUES))
return -(GaussianLOTHypothesis.compute_likelihood(self, data) + constant_prior)
def compute_prior(self):
self.prior = GaussianLOTHypothesis.compute_prior(self)
self.prior += sum(map(lambda x: normlogpdf(x,0.0,CONSTANT_SD), self.CONSTANT_VALUES))
self.posterior_score = self.prior + self.likelihood
return self.prior
def compute_prior(self):
# Add together the LOT prior and the constant prior, here just a gaussian
return LOTHypothesis.compute_prior(self) +\
sum(map(lambda x: normlogpdf(x,0.0,self.constant_sd), self.parameters))
def compute_prior(self):
# Add together the structural prior and the constant prior
return LOTHypothesis.compute_prior(self) +\
sum(map(lambda x: normlogpdf(x,0.0,self.constant_sd), self.CONSTANT_VALUES))
